[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: working with fixed point

From: Per Persson
Subject: Re: working with fixed point
Date: Mon, 25 Aug 2003 10:31:38 +0200

On Monday, August 25, 2003, at 11:56 AM, Zdenek Hurak wrote:

Well, I think this will not be as easy as that. Converting the data is one step only, but Ravid obviously wants to calculate in fixed point arithmetics

multiplying two integers will give the same result whether or not that multiplication was performed using integer or floting point arithmetics. With a limited number of bits available for representation, the result may overflow. This will have to be handled in some way, such as truncating or saturating the result, in the fpga (or VHDL/Verilog code) as well as in an octave script.

It is quite possible to have a per-sample correspondance between the output of a VHDL simulation and a simulation run in octave.

The following is an example of a first order link of a wave digital filter. It was used to simulate the effect of varying wordlength in octave before moving to VHDL code and simulation.

function [y state] = qaplink1(x, alpha, B, W, state);
u = alpha*(-x+state);
## Quantize u
u = truncate(u, B, W-1);
y = u + state;
y = saturate(y, B);
state = u + x;
state = saturate(state, B);
## endfunction

Considering that the "fixed-point skills" are by far the most expensive
component in commercial Matlab suite, it can be expected that solution is not

AFAIK commercial software is not normally priced by its complexity or the amount of time spent developing it, but rather from an estimate of how much that particular functionality is worth to the customer.


Per Persson
Blekinge Institute of Technology
Dept. of Signal Processing and Telecommunications

e-mail: address@hidden

Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:
How to fund new projects:
Subscription information:

reply via email to

[Prev in Thread] Current Thread [Next in Thread]